Interpreting and Exploiting Functional Specialization in Multi-Head Attention under Multi-task Learning
This work addresses the interpretability and efficiency of transformer models in multi-task learning, offering a method to reduce negative transfer and boost performance, though it is incremental in nature.
The study investigated whether multi-head attention in transformers develops functional specialization during multi-task training, similar to the human brain, and found that it does, influenced by task similarity. They introduced a training method to enhance this specialization, which improved performance in multi-task and transfer learning without extra parameters.
Transformer-based models, even though achieving super-human performance on several downstream tasks, are often regarded as a black box and used as a whole. It is still unclear what mechanisms they have learned, especially their core module: multi-head attention. Inspired by functional specialization in the human brain, which helps to efficiently handle multiple tasks, this work attempts to figure out whether the multi-head attention module will evolve similar function separation under multi-tasking training. If it is, can this mechanism further improve the model performance? To investigate these questions, we introduce an interpreting method to quantify the degree of functional specialization in multi-head attention. We further propose a simple multi-task training method to increase functional specialization and mitigate negative information transfer in multi-task learning. Experimental results on seven pre-trained transformer models have demonstrated that multi-head attention does evolve functional specialization phenomenon after multi-task training which is affected by the similarity of tasks. Moreover, the multi-task training strategy based on functional specialization boosts performance in both multi-task learning and transfer learning without adding any parameters.